A number of reports have appeared, most by one person, that purport to show an ability to identify cancer patients by specific handwriting characteristics, presumably evidence of neuromuscular dysfunction associated with cancer. The claimed power of discrimination was very high. Some questions about the possibility of non-blind experimentation leave a cloud over the validity of the reported findings. This work is intended to test the claim with rigorous, blind analysis of handwriting samples and blind statistical analysis. Possible differences in capability of discrimination for different sites will be examined.